summaryrefslogtreecommitdiffstats
path: root/Demo/embed/README
diff options
context:
space:
mode:
authorGuido van Rossum <guido@python.org>2001-01-10 17:11:51 (GMT)
committerGuido van Rossum <guido@python.org>2001-01-10 17:11:51 (GMT)
commit3559d1f9b3f33b45de924f5d17dcea6ab949bb41 (patch)
treeaa46bef878d40f1ad03d2b7e78bd980e897fbbda /Demo/embed/README
parent7339f4c72d2eaf9fef2416d3328207dcc655f578 (diff)
downloadcpython-3559d1f9b3f33b45de924f5d17dcea6ab949bb41.zip
cpython-3559d1f9b3f33b45de924f5d17dcea6ab949bb41.tar.gz
cpython-3559d1f9b3f33b45de924f5d17dcea6ab949bb41.tar.bz2
Add loop.c -- a test program for repeatedly calling Py_Initialize()
and Py_Finalize(). It seems to dump core right now...
Diffstat (limited to 'Demo/embed/README')
-rw-r--r--Demo/embed/README7
1 files changed, 7 insertions, 0 deletions
diff --git a/Demo/embed/README b/Demo/embed/README
index 8858c71..a0f7af8 100644
--- a/Demo/embed/README
+++ b/Demo/embed/README
@@ -10,3 +10,10 @@ the source directory (../..)
2) change the variables that together define the list of libraries
(MODLIBS, LIBS, SYSLIBS) to link with, to match their definitions in
$(blddir)/Modules/Makefile
+
+An additional test program, loop.c, is used to experiment with memory
+leakage caused by repeated initialization and finalization of the
+interpreter. It can be build by saying "make loop" and tested with
+"make looptest". Command line usage is "./loop <python-command>",
+e.g. "./loop 'print 2+2'" should spit out an endless number of lines
+containing the number 4.